Posted by Dan Guzman on 06/06/07 12:19

> So, I'll be forced to have 3077 filegroups, grouped with 50
> partitions.
> Is there a way to have a partition function/scheme that sees other
> schemes, instead of filegroups?

Why do you need separate filegoups? It seems to me that the main purpose of
partitioning here is for manageability and all those files/filegroups only
add to administration complexity and wasted space.

You might consider a hybrid solution with 50 individual state tables
included in a partitioned view, with each state table partitioned by county.
This approach would leverage partitioning to quickly reload individual
counties yet provide a seamless view of the entire country.


--
Hope this helps.

Dan Guzman
SQL Server MVP

"Piero 'Giops' Giorgi" <giorgi.piero@gmail.com> wrote in message
news:1181087950.821544.175350@j4g2000prf.googlegroups.com...
> On Jun 5, 3:24 pm, Erland Sommarskog <esq...@sommarskog.se> wrote:
>
>>> But, before trying, can I have 3077 files in ONE partition, and drop
>>> all the states stuff?
>
>> No, in the topic for CREATE PARTITION FUNCTION, I found that you
>> cannot have more than 999 boundary values.
>
> Dang it... I kew there was a catch.
>
> So, I'll be forced to have 3077 filegroups, grouped with 50
> partitions.
> Is there a way to have a partition function/scheme that sees other
> schemes, instead of filegroups?
>
> I mean Filegroups Counties (3077) - grouped by state (50) - all
> together in ONE partitioned table.
>
> Any Ideas?
>
> Thank you!
>
> Piero
>

[Back to original message]